Novel Rapid Test for Detecting Carbapenemase

Yanfang Feng,1 Akilan Palanisami,1 Jerrin Kuriakose, Michael Pigula, Shoaib Ashraf, Tayyaba HasanAuthor affiliations: Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA (Y. Feng, A. Palanisami, J. Kuriakose, M. Pigula, S. Ashraf, T. Hasan); Harvard-MIT Health Sciences and Technology, Cambridge, Massachusetts, USA (T. Hasan)

DOI: https://doi.org/10.3201/eid2604.181655

As a potent β-lactamase, carbapenemase can de-grade almost all β-lactam antimicrobial drugs,

including the carbapenems, regarded as the last line of therapy for many life-threatening infections (1,2). Various epidemic types of carbapenemase have been reported globally, including Klebsiella pneumoniae carbapenemase, Verona integron-encoded metallo-β-lactamase, Serratia marcescens enzyme, imipen-em-hydrolyzing β-lactamase, New Delhi metallo-β-lactamase, oxacillinase, metallo-β-lactamase, and São Paulo metallo-β-lactamase (1). If uncontrolled, the spread of these carbapenemases is expected to increase therapeutic failure and leave many patients with no effective treatment options.

Despite the urgency, timely carbapenemase de-tection remains a challenge for microbiology laborato-ries. Phenotypic assays are inexpensive and easily per-formed, but their use requires 24–48 hours and many lack sensitivity or specificity (3). The widespread use of other assays (e.g., molecular tests of carbapenemase genes, mass spectrometry detection of carbapenem hydrolysis) is impeded by the expertise required to perform them and their cost (4,5). The recently devel-oped (2012) Carba NP test and variants are elegant solutions, but their use requires up to 2 hours (6). Further improvements in test rapidity and simplicity are highly desirable, especially for patients in critical condition, who need immediate therapy and infection control action.

We demonstrate that by using fluorescence identification of β-lactamase activity (FIBA), car-bapenemase production in bacteria can be detected

sensitively and specifically in 10 minutes, with only 1 step. FIBA uses a dark fluorescence probe, β-LEAF (β-lactamase enzyme–activated fluoro-phore), which turns fluorescent when cleaved by β-lactamases, including penicillinases, extended-spectrum β-lactamases (ESBL), AmpC β-lactamases, and carbapenemases (7,8). Thus, the rate of fluo-rescence increase (hereafter called increase rate) is a measure of the bacterial β-lactamase activity and is reduced as the β-lactamase activity is ham-pered. For a noncarbapenemase β-lactamase, the increase rate will be reduced by the addition of imipenem, which binds the enzyme active site and blocks β-LEAF access (2). In contrast, the increase rate for a carbapenemase is relatively unaffected by imipenem addition because carbapenemase is able to rapidly cleave the imipenem and relieve the inhibition (1). Accordingly, bacteria that produce carbapenemases can be detected by comparing the increase rate with and without imipenem (Figure; Appendix Figure 1, https://wwwnc.cdc.gov/EID/article/26/4/18-1655-App1.pdf).

FIBA is performed in a 96-well plate. Each well contains 50 µL of 20 µmol/L β-LEAF, 25 µL of phos-phate-buffered saline with or without 40 µmol/L imipenem (Cayman Chemical, https://www.cay-manchem.com), and 10 µL of either 1 mg/mL poly-myxin B nonapeptide or 1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (Sigma-Al-drich, https://www.sigmaaldrich.com), which act as weak or strong permeabilizers, respectively. To start the assay, 25 µL of 1 × 1010 CFU/mL bacte-rial suspension made by colonies grown overnight on BHI agar (Sigma-Aldrich) is added to each well. To monitor the increase rate, fluorescence measure-ment is then performed at 37°C at 10-s intervals for 10 min with Ex/Em 450/510 nm in the plate reader (Spectramax M5 plate reader, Molecular Devices, https://www.moleculardevices.com). For each bacterial sample, we performed the reactions in duplicate and averaged the results. We objectively interpreted the fluorescence measurements by us-ing an automated Python script (Appendix), which required a few seconds after assay completion.

We tested FIBA on 76 randomly selected in-fection isolates from either the Centers for Disease Control and Prevention (9) or the American Type Culture Collection (https://www.atcc.org). The MICs of these isolates, if not predetermined, were measured by the 2017 Clinical Laboratory and Stan-dards Institute (https://clsi.org) broth dilution method. Genetic test results for β-lactam resistance were provided with the isolates. Among these, 55

793 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 26, No. 4, April 2020

RESEARCH LETTERS

1These authors contributed equally to this article.

We developed a carbapenemase test based on the ability of imipenem to inhibit noncarbapenemase β-lactamases. The test uses bacterial isolates with a fluorescent β-lactamase substrate, producing objective results with 100% sensitivity and specificity in 10 minutes. The assay is inexpensive and consists of only 1 mixing step.

were carbapenemase positive, carrying the major epidemic carbapenemase types including K. pneu-moniae carbapenemase (n = 20), imipenem-hydro-lyzing β-lactamase (n = 2), metallo-β-lactamase (n = 4), New Delhi metallo-β-lactamase (n = 10), oxa-cillinase (n = 8), S. marcescens enzyme (n = 2), São Paulo metallo-β-lactamase (n = 1), Verona integron-encoded metallo-β-lactamase (n = 6), and New Delhi metallo-β-lactamase oxacillinase (n = 2). The other 21 isolates expressed noncarbapenemase β-lactamases, which involved 9 isolates with only ESBL, 3 isolates with both ESBL and porin modification, 6 isolates with only AmpC β-lactamase, and 3 isolates with both ESBL and AmpC β-lactamase. Among these isolates, 3 were carbapenem resistant. The entire panel, which included 28 colistin-resistant strains (MIC >4 ug/mL), was classified successfully with FIBA (Appendix Tables 1, 2), resulting in 100% sen-sitivity (95% CI 94%–100%) and 100% specificity (95% CI 84%–100%).

The primary limitation of this study is the small number of isolates evaluated. However, the breadth of isolates studied here included 8 enzyme types across 16 species, suggesting the generality of the approach.

FIBA can be performed ≈10 times faster than the most rapid carbapenemase test commercially avail-able while maintaining comparable sensitivity and specificity (6,10). Its automated analysis improves turnaround time and reduces operator variability. With a reagent cost/assay of ≈US $1, FIBA is close in price to phenotypic tests but substantially faster and less labor intensive. Furthermore, the FIBA paradigm is extensible; by replacing imipenem with other known subtype-dependent inhibitors of carbapenemase (e.g., clavulanic acid, EDTA), rapid carbapenemase subtyping may also be possible. Our study demonstrates that low-cost, rapid assess-ment of carbapenemase can be performed in a 1-step format suitable for large-scale epidemiologic stud-ies, thereby providing a new tool for infection out-break control.

AcknowledgmentsWe acknowledge Gerry Nau for helpful conversations.

The research was supported by NIH/FIC R21TW010202 and DOD/AFOSR FA9550-16-1-0479. Patent pending for the β-LEAF design (application no. US20180094292A1).

794 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 26, No. 4, April 2020

RESEARCH LETTERS

Figure. Schematic illustration of the principle of fluorescence identification of β-lactamase activity. A). The β-lactamase–activated fluorophore probe comprises a cleavable β-lactam core conjugated to 2 fluorophores (circled) that are quenched because of close proximity. This construct was designed to mimic the enzymatic degradation properties of easily cleavable β-lactam antimicrobial drugs. When this probe is attacked by β-lactamase, the probe core is cleaved, leading to the separation of the fluorophores and the recovery of their fluorescent properties (fluorescent state). B) Assay profile for carbapenemase-producing bacteria. C) Assay profile for non–carbapenemase-producing bacteria. Black, quenched fluorophore; blue, unquenched fluorophore turning fluorescent; green, β-lactam core; red, imipenem; purple, β-lactamase. β-LEAF, β-lactamase enzyme–activated fluorophore.

About the AuthorDr. Feng is a research fellow at Massachusetts General Hospital and Harvard Medical School, Boston. Her major research interests are molecular mechanisms, epidemiology, and prevention of drug resistance in bacteria.

References 1. Queenan AM, Bush K. Carbapenemases: the versatile

beta-lactamases. Clin Microbiol Rev. 2007;20:440–58. https://doi.org/10.1128/CMR.00001-07

2. Papp-Wallace KM, Endimiani A, Taracila MA, Bonomo RA. Carbapenems: past, present, and future. Antimicrob Agents Chemother. 2011;55:4943–60. https://doi.org/10.1128/AAC.00296-11

3. Tamma PD, Opene BN, Gluck A, Chambers KK, Carroll KC, Simner PJ. Comparison of 11 phenotypic assays for accurate detection of carbapenemase-producing Enterobacteriaceae. J Clin Microbiol. 2017;55:1046–55. https://doi.org/10.1128/JCM.02338-16

4. Lasserre C, De Saint Martin L, Cuzon G, Bogaerts P, Lamar E, Glupczynski Y, et al. Efficient detection of carbapenemase activity in Enterobacteriaceae by matrix- assisted laser desorption ionization–time of flight mass spectrometry in less than 30 minutes. J Clin Microbiol. 2015;53:2163–71. https://doi.org/10.1128/JCM.03467-14

5. Monteiro J, Widen RH, Pignatari AC, Kubasek C, Silbert S. Rapid detection of carbapenemase genes by multiplex real-time PCR. J Antimicrob Chemother. 2012;67:906–9. https://doi.org/10.1093/jac/dkr563

6. Nordmann P, Poirel L, Dortet L. Rapid detection of carbapenemase-producing Enterobacteriaceae. Emerg Infect Dis. 2012;18:1503–7. https://doi.org/10.3201/eid1809.120355

7. Sallum UW, Zheng X, Verma S, Hasan T. Rapid functional definition of extended spectrum β-lactamase activity in bacterial cultures via competitive inhibition of fluorescent substrate cleavage. Photochem Photobiol. 2010;86:1267–71. https://doi.org/10.1111/j.1751-1097.2010.00801.x

8. Khan S, Sallum UW, Zheng X, Nau GJ, Hasan T. Rapid optical determination of β-lactamase and antibiotic activity. BMC Microbiol. 2014;14:84. https://doi.org/10.1186/ 1471-2180-14-84

9. CDC and FDA Antibiotic Resistance Isolate Bank [cited 2019 Mar 8]. https://wwwn.cdc.gov/ARIsolateBank

10. Maurer FP, Castelberg C, Quiblier C, Bloemberg GV, Hombach M. Evaluation of carbapenemase screening and confirmation tests with Enterobacteriaceae and development of a practical diagnostic algorithm. J Clin Microbiol. 2015;53:95–104. https://doi.org/10.1128/JCM.01692-14

Address for correspondence: Tayyaba Hasan, The Wellman Center for Photomedicine, Massachusetts General Hospital, 40 Blossom St, Boston, MA 02114, USA; email: thasan@mgh.harvard.edu

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 26, No. 4, April 2020 795

RESEARCH LETTERS

Clonal complex 398 methicillin-resistant Staphylococcus aureus (MRSA) is a typical lineage of livestock-associ-ated MRSA. We report a case of intractable arthritis of the shoulder joint caused by a multidrug-resistant Pan-ton-Valentine leukocidin–positive livestock-associated MRSA clonal complex 398 sequence type 1232 clone in a patient in Japan who had no animal contact.

Arthritis Caused by MRSA CC398 in Patient without Animal Contact, Japan

Hidemasa Nakaminami,1 Yuji Hirai,1 Hirosuke Nishimura, Shunsuke Takadama, Norihisa NoguchiAuthor affiliations: Tokyo University of Pharmacy and Life Sciences, Tokyo, Japan (H. Nakaminami, S. Takadama, N. Noguchi); Tokyo Medical University Hachioji Medical Center, Tokyo (Y. Hirai, H. Nishimura)

DOI: https://doi.org/10.3201/eid2604.190376

In the past decade, methicillin-resistant Staphylo-coccus aureus (MRSA) has been detected in live-

stock, including swine, poultry, and veal calves (1,2). In general, the virulence of animal-derived livestock-associated (LA-MRSA) strains is consid-ered to be lower than that of community-acquired MRSA lineages (3). However, LA-MRSA strains can effectively colonize and infect humans, with subsequent transmission in both community and hospital settings. Human colonization with LA-MRSA sequence type (ST) 398 was first recognized among swine farmers in France and the Nether-lands in the early 2000s (4). According to Larsen et al., clonal complex (CC) 398 MRSA accounted for 21% of MRSA isolated from skin and soft tissue infections in Denmark during 2010–2015 (5). How-ever, ST398 MRSA has not been isolated in patients in Japan. We report a case of intractable arthritis of the shoulder joint caused by a multidrug-resistant Panton-Valentine leukocidin (PVL)–positive LA-MRSA CC398 (ST1232) clone in a patient in Japan who had no animal contact.

We performed MRSA identification, staphylococ-cal cassette chromosome (SCC) mec typing, spa typing, multilocus sequence typing (MLST), MIC determina-tion, and PCR assays for detecting virulence factors and antimicrobial resistance genes, as described pre-viously (1,6). The study protocol was approved by the Tokyo University of Pharmacy and Life Sciences Ethics Committee (approval no. 12–09).

1These authors contributed equally to this article.

Page 1 of 5

Article DOI: https://doi.org/10.3201/eid2604.181655

Novel Rapid Test for Detecting Carbapenemase

Appendix

Supplementary Methods

The imipenem solution was prepared and used within 30 mins (stored at 4 °C during this

time). Samples were physically tested in a blind and random fashion. To develop the subsequent

analysis, the data were unblinded to determine the appropriate analysis thresholds. The FIBA

limit of detection was 109-10 CFU/ml for this panel.

The automated Python analysis for FIBA test

To quickly determine carbapenemase activity from the fluorescence time course data, a

python script was developed to easily analyze the excel spreadsheet files generated by the

fluorescence plate reader. The analysis is as follows: To quantify the changes of the β-lactamase

activity upon the addition of imipenem, a β-lactamase inhibition index (BI, see equation II),

defined as the ratio of fluorescence increase rate (R, see equation I) between wells without and

with imipenem, was created. BI increases with imipenem inhibition (non-carbapenemase β-

lactamase behavior), and an isolate with BI≤0 is classified as carbapenemase-positive.

Among the samples with BI>0 in the presence of 100 μg/ml cell membrane permeabilizer

polymyxin B nonapeptide (PMBN), a parallel assay with another permeabilizer, 0.1% 3-[(3-

cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), was also performed. This is

to rule out false negatives that may be caused by insufficient permeabilization due to bacterial

polymyxin resistance. Isolates with a BI≤0 under the condition of CHAPS were also classified as

carbapenemase-positive. All remaining isolates were classified as carbapenemase-negative.

The equations used in Python analysis:

I.

Page 2 of 5

𝐑𝐑 = �𝑣𝑣𝑖𝑖

𝑁𝑁

𝑖𝑖=1

where v is the fluorescence increase per time step, i is the time index of the sorted v (i.e., v1 is the

largest fluorescent difference, v2 is the second largest fluorescent difference, etc.) and N is the

number of v to be summed. v was determined from a 50 second running average of the

fluorescent time course, and N was set to 45.

II.

𝑩𝑩𝑩𝑩 = ��𝑣𝑣𝑖𝑖0 𝜇𝜇𝜇𝜇

𝑁𝑁

𝑖𝑖=1

�𝑣𝑣𝑖𝑖10 𝜇𝜇𝜇𝜇

𝑁𝑁

𝑖𝑖=1

� � − BI0

where 𝑣𝑣𝑖𝑖𝐶𝐶 is the fluorescence increase per time step at imipenem concentration 𝐶𝐶 , and BI0,

designated as the BI cut-off value for the enzyme inhibition, is 1.25.

Appendix Table 1. Carbapenemase-producing isolates subjected to the FIBA* test Carbapenemase category

Species No.† Carbapenem susceptibility‡ FIBA test

result§ Ambler class Type Subtype Imi Mer Ert Dor A KPC KPC-2 C.freundii 1 16 > 8 > 8 > 8 +

E.cloacae 1 8 8 ˃ 8 4 + K.pneumoniae 1 ˃ 8 16 ≤ 0.25 ˃ 8 +

M.morganii 1 8 4 8 4 + P.mirabilis 1 16 0.5 1 4 +

KPC-3 E.cloacae 2 ≥8 4-8 0.5 1- > 8 + E.coli 2 4-8 4 0.5 - 8 ≤ 0.25 - 4 +

K.ascorbata 1 4 8 8 4 + K.oxytoca 1 4 1 0.5 0.5 + K.ozaenae 1 ˃ 8 ˃ 64 0.5 ˃ 8 +

K.pneumoniae 3 ˃ 8 ˃ 8 ≥ 4 ˃ 8 + R.ornithinolytica 1 4 1 1 2 +

KPC-4 E.cloacae 1 0.5 1 0.5 ˃ 8 + KPC-5 P.aeruginosa 1 ˃ 8 ˃ 64 1 ˃ 8 + KPC-6 E.cloacae 1 4 4 0.5 8 +

P.mirabilis 1 16 2 2 2 + SME SME-3 S.marcescens 2 ˃ 8 ˃ 64 4 ≤ 0.25- 0.5 +

NMC-A NMC-A E.cloacae 2 ≥ 32 > 8 > 8 > 8 + B NDM NDM-1 E.coli 1 8 > 16 > 8 > 8 +

K.pneumoniae 2 ˃ 8 > 8 1 - > 8 > 8 + M.morganii 1 2 8 4 ˃ 8 + P.mirabilis 1 32 4 4 > 8 + P.rettgeri 1 8 32 4 ≤ 0.25 +

S.senftenberg 1 4 8 > 8 8 + A.baumannii 1 ˃ 8 64 1 > 8 +

Citrobacter spp. 1 16 > 8 > 8 > 8 + E.cloacae 1 4 32 ≤ 0.25 ˃ 8 +

NDM-1/OXA-64 A.baumannii 1 ˃ 8 64 0.5 > 8 + NDM-6 E.coli 1 16 > 8 > 8 > 8 +

VIM VIM-1 E.cloacae 1 4 2 2 4 + K.pneumoniae 1 4 4 1 4 +

VIM-2 P.aeruginosa 2 ˃ 64 ˃ 8 ˃ 8 ˃ 8 + VIM-27 K.pneumoniae 1 64 ˃ 8 ˃ 8 ˃ 8 + VIM-4 P.aeruginosa 1 ˃ 64 ˃ 8 ˃ 8 4 +

IMP IMP-1 P.aeruginosa 1 ˃ 64 ˃ 8 ˃ 8 ˃ 8 + IMP-14 P.aeruginosa 1 64 ˃ 8 ˃ 8 ˃ 8 + IMP-4 K.pneumoniae 2 1 - 4 2 - 4 2 - 4 4 - 8 +

Page 3 of 5

Carbapenemase category Species No.†

Carbapenem susceptibility‡ FIBA test result§ Ambler class Type Subtype Imi Mer Ert Dor

SPM SPM-1 P.aeruginosa 1 ˃ 64 ˃ 8 ˃ 8 ˃ 8 + D OXA OXA-48 E.aerogenes 1 4 2 2 2 +

K.pneumoniae 1 4 8 ˃ 8 8 + OXA-58/100 A.baumannii 2 16 - 32 ˃ 8 ˃ 8 8 + OXA-66/72 A.baumannii 1 ˃ 64 ˃ 8 ˃ 8 ˃ 8 + OXA-181 K.ozaenae 1 4 4 ˃ 8 4 +

K.pneumoniae 1 2 4 ˃ 8 4 + OXA-232 K.pneumoniae 1 4 ˃ 8 > 8 ˃ 8 +

*FIBA, Fluorescence identification of β-Lactamase activity. †No., number of isolates tested. ‡MIC of the tested isolates for doripenem (Dor), ertapenem(Ert), imipenem(Imi) and meropenem(Mer). §FIBA test result: -, negative; +, positive. All the results shown here were based on the average of two independent replicates. With the permeabilizer PMBN, there are 4 out of 57 (7%) isolates labeled as false negatives which are subsequently found positive with the permeabilizer CHAPS.

Appendix Table 2. Non-carbapenemase-producing isolates subjected to the FIBA* test

β-Lactamase category Species No.†

Carbapenem susceptibility‡ FIBA test result§ Type Subtype Imi Mer Ert Dor

ESBL CTX-M-14, TEM-1B E.coli 1 ≤ 0.5 ≤ 0.12 1 ≤ 0.12 - CTX-M-15, SHV-1,

TEM-1B, OXA-1/10¶ K.pneumoniae 1 1 2 > 8 2 -

CTX-M-2, SHV-83,

TEM-1A, OXA-9/10¶ K.pneumoniae 1 8 > 8 > 8 > 8 -

SHV-3 E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 - SHV-4 E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 - SHV-12 K.pneumoniae 1 ≤ 0.5 2 > 8 1 -

SHV-12¶ K.pneumoniae 1 ≤ 0.5 ≤ 0.12 0.25 ≤ 0.12 - TEM-3 E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 - TEM-10 E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 - TEM-12 E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 - TEM-26 E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 -

TEM-52B E.coli 1 ≤ 0.5 0.25 1 0.25 - AmpC ACT-7 E.cloacae 1 ≤ 0.5 ≤ 0.12 0.25 ≤ 0.12 -

ACT-15 E.cloacae 1 ≤ 0.5 ≤ 0.12 1 ≤ 0.12 - CMY-2 E.coli 2 ≤ 0.5 ≤ 0.12 - 1 ≤ 0.12 - 2 ≤ 0.12 - cAmpC E.aerogenes 1 ≤ 0.5 ≤ 0.12 1 ≤ 0.12 - cAmpC E.cloacae 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 -

ESBL& AmpC

CTX-M14; DHA-1, SHV-11, TEM-1B¶

K.pneumoniae 1 16 8 > 8 8 -

cAmpC, TEM-1B E.cloacae 1 ≤ 0.5 ≤ 0.12 1 ≤ 0.12 - CMY-2, TEM-1B E.coli 1 ≤ 0.5 ≤ 0.12 ≤ 0.12 ≤ 0.12 - *FIBA, fluorescence identification of β-Lactamase activity. †No., number of isolates tested. ‡MIC of the tested isolates for doripenem (Dor), ertapenem(Ert), imipenem(Imi) and meropenem(Mer). §FIBA test result: -, negative; +, positive. ¶Porin modifications were present together with β-lactamase. All the results shown here were based on the average of two independent replicates.

Page 4 of 5

Appendix Figure 1. The detection of bacterial carbapenemase production by the fluorescence

identification of β-lactamase activity. A) Carbapenemase producing isolates cleave β-lactamase enzyme-

activated fluorophore irrespective of imipenem addition, as exemplified here by the strain # 0147 from the

CDC isolate bank. B) Non-carbapenemase β-lactamases are unable to cleave β-lactamase enzyme-

activated fluorophore when inhibited by imipenem, as shown here by the isolate # 0065 from the CDC

isolate bank.

Page 5 of 5

Appendix Figure 2. The rate of fluorescence increase in FIBA* increases with the addition of PMBN†.

*FIBA, fluorescence identification of β-lactamase activity; †PMBN, polymyxin B nonapeptide; The strain

used here as an illustration is a β-lactamase producing strain from ATCC (Escherichia coli, ATCC® BAA-

196™).